Recent publications have found an association between variants of exostosin 2 (EXT2) gene and the risk of type 2 diabetes in some population but not in others. In an attempt to address these inconsistencies, we investigated EXT2 variants in two independent cohorts, and conducted a literature-based meta-analysis. Through regression model, we assessed the relationship between the EXT2 single nucleotide polymorphisms (SNPs) (rs3740878, rs11037909 and rs1113132) and the risk of type 2 diabetes in Han Chinese population, including a total of 2,533 cases and 2,643 controls. We combined our data with that from previously published studies and performed a meta-analysis to evaluate the effect size of the gene. Consistent with some studies, we found marginal association for the rs3740878 (OR = 1.07, 95 % CI = 0.99, 1.16, p = 0.09), rs11037909 (OR = 1.07, 95 % CI = 0.99, 1.16, p = 0.08), and rs1113132 (OR = 1.08, 95 % CI = 1.00, 1.17, p = 0.06) in our 2 cohorts. Meta-analysis, comprising 9,224 type 2 diabetes and 10,484 controls, revealed that three SNPs (rs3740878, rs11037909 and rs1113132) in EXT2 were significantly associated with type 2 diabetes (ORs range from 1.06 to 1.07, p = 0.038, p = 0.008 and p = 0.005, respectively). Variation in the EXT2 locus appears to be associated with a small increase in the risk of type 2 diabetes. However, well-designed prospective studies with larger sample size and more ethnic groups are needed to further validate the results.